Skip to main content

Table 5 PESQ values for the W β-SA estimator with different schemes of parameter β, case of babble noise

From: Bayesian STSA estimation using masking properties and generalized Gamma prior for speech enhancement

Input SNR (dB) −10 −5 0 5 10
Input noisy speech 1.31 1.56 1.83 2.14 2.43
Choice of β = 1.82 1.49 1.73 2.04 2.42 2.73
Choice of β by (11) 1.55 1.88 2.18 2.46 2.76
Choice of β by (12) 1.55 1.88 2.17 2.47 2.79
Proposed choice of β 1.58 1.91 2.23 2.51 2.82